Proving partial order

162 Views Asked by At

In few days I'm having an exam from Formal Logic and Set Theory and would like to confirm if my method of proving partial orders is correct. The task:

For $(n, m), (p, q) \in \mathbb{N}_{+}^{2}$, $$(n, m) \preceq (p, q) \Leftrightarrow (n, m) = (p, q) \vee (n < p \wedge n - m \leq p - q).$$ Prove that $\preceq$ is a partial order.

Here is my idea for the proof:

1) Reflexive

(a, b) $ \preceq $ (a, b)

$(a, b) = (a, b) \vee (a < a \wedge (a - b) \leq (a - b))$

The first part is true, so it's reflexive.

2) Antisymmetric

$ (a, b) \preceq (c, d) \wedge (c, d) \preceq (a, b) \Rightarrow (a, b) = (c, d) $

My idea was to write above condition using actual rules given in the task:

$ [(a, b) = (c, d) \vee ((a < c) \wedge (a - b) \leq (c - d))] \wedge [(c, d) = (a, b) \vee ((c < a) \wedge (c - d) \leq (a - b))] $

If we mark every part in brackets divided with $ \vee $ with letters p, q, r, s, we can see that distribution of p and r will give us:

$ [(a, b) = (c, d) \wedge (c, d) = (a, b))] $

And that's what we were looking for. So antisymmetry is true.

3) Transitive

$ (a, b) \preceq (c, d) \wedge (c, d) \preceq (e, f) \Rightarrow (a, b) \preceq (e, f) $

I used the similar procedure to what I did above. Mainly I wrote the conditions:

$ [(a, b) = (c, d) \vee ((a < c) \wedge (a - b) \leq (c - d))] \wedge [(c, d) = (e, f) \vee ((c < e) \wedge (c - d) \leq (e - f))] $

I once again used p, q, r, s notation and distributed. Again, distribution of p and r gives us:

$ [(a, b) = (c, d) \wedge (c, d) = (e, f)] $

Which should prove the last part, because that's the condition of partial order in the task. Is that the correct way of proving partial orders? I will be very thankful for tips and answers, because they will help me pass the course.

1

There are 1 best solutions below

3
On BEST ANSWER

I would use a different approach in your proofs of anti-symmetry and transitivity.
This way, you don't need distribution.

Anti-symmetry. The second branch of the definitions doesn't apply, since that would include $a<c$ and $c<a$, a contradiction.
Thus, you get immediatelly $(a,b) = (c,d)$ (by definition), which is what was to be proven.

Transitivity. It's a similar situation.
If $(a,b) = (c,d)$, then $(a,b) \leq (e,f)$ because $(c,d) \leq (e,f)$; the same happens if $(c,d) = (e,f)$.
So the only non-trivial situation is if $$a<c<e$$ and $$a-b < c-d < e-f,$$ from which it follows that $a < e$ and $a-b< e-f$ (it turns out to be rather trivial too).